Electron microscopy studies of lithium aluminium hydrides

Article abstract of DOI:10.1016/j.jallcom.2004.11.058

The microstructure of LiAlD₄ has been studied during the decomposition process using powder X-ray diffraction (PXD), transmission electron microscopy (TEM) and scanning electron microscopy (SEM). Selected area diffraction (SAD) confirmed the ex

Full text of DOI:10.1016/j.jallcom.2004.11.058

Journal of Alloys and Compounds 395 (2005) 307–312
Electron microscopy studies of lithium aluminium hydrides
C.M. Andreia, J. Walmsleyb, D. Blanchardc, H.W. Brinksc, R. Holmestada,∗, B.C. Haubackc
a
Department of Physics, Norwegian University of Science and Technology, NO-7491Trondheim, Norway
b
SINTEF Materials and Chemistry, NO-7465 Trondheim, Norway
c
Institute for Energy Technology, P.O. Box 40, NO-2027 Kjeller, Norway
Received 5 November 2004; accepted 18 November 2004
Available online 7 January 2005
Abstract
The microstructure of LiAlD4 has been studied during the decomposition process using powder X-ray diffraction (PXD), transmission
electron microscopy (TEM) and scanning electron microscopy (SEM). Selected area diffraction (SAD) confirmed the existence of all the
decomposed phases presented in the X-ray powder diffraction data. High-resolution transmission electron microscopy confirmed the presence
of the LiAlD4 phase before decomposition. After the first decomposition reaction, the electron diffraction patterns were dominated by metallic
Al, but significant amounts of Li3AlD6 were also present. Inhomogeneous distribution of the phases was confirmed by scanning electron
microscopy. Most of the Al was present in discrete particles on the surface of the decomposed alanate. For the LiAlD4 sample heated to
275 ◦C under dynamical vacuum, many of the interplanar distances for Al and LiD phase overlap. Some of the Li3AlH6 phase was found to
remain in the material and two characteristic spacings were identified in both the electron and X-ray diffraction data. Fine scale variation in
composition was observed in the backscattered scanning electron image.
© 2004 Elsevier B.V. All rights reserved.
Keywords: LiAlH4; Alanates; Complex hydride; Electron microscopy; Diffraction pattern; TEM; SEM
1. Introduction
moderate conditions. However, it is interesting to consider
the details of the decomposition process in the material. The
hydrogen storage capacity of LiAlH4 is very high, 10.6 wt.%
in total. Decomposition takes place in three steps upon heat-
ing [11–14] with 5.3, 2.6 and 2.6 wt.% release of hydrogen,
respectively.
Hydrogen storage has been the subject of intensive re-
search in recent years. Hydrogen is a promising alternative
fuelsinceitispollution-freeandcanreadilybeproducedfrom
renewable energy resources [1]. Complex hydrides contain-
ing aluminium (alanates) are attractive as hydrogen storage
compounds due their high hydrogen content and low weight.
Alanates, such as, NaAlH4 and LiAlH4, have been known
for a long time, but the work of Bogdanovic and Schwickardi
[2] who demonstrated reversible hydriding behaviour in
NaAlH4 when Ti based additive is added, have brought these
complex hydrides into focus for storage applications.
3LiAlH4 = Li3AlH6 + 2Al + 3H2
(R1)
(R2)
3
Li3AlH6 = 3LiH + Al + H2
2
3
2
3LiH + 3Al = 3LiAl + H2
(R3)
The last reaction occurs above 400 ◦C and is not consid-
eredavailableforpracticalpurposes. Thedecompositiontem-
peratures are reported to be between 150 and 175 ◦C for re-
action 1 (R1) and 180–220 ◦C for reaction 2 (R2) [13]. The
temperatures can be reduced to 112 ◦C for R1 and 127 ◦C for
R2 by using very slow heating rates [11].
Since the first synthesis of LiAlH4 in 1947 [3] several
groups have further studied structural and desorption prop-
erties of this system [4–11]. LiAlH4 is not reversible under
∗
Electron microscopy can provide key information about
the microstructure of the alanates, which has not been
Corresponding author. Tel.: +47 73 59 38 80; fax: +47 73 59 77 10.
E-mail address: randi.holmestad@phys.ntnu.no (R. Holmestad).
0925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2004.11.058

308
C.M. Andrei et al. / Journal of Alloys and Compounds 395 (2005) 307–312
accessible by other techniques. However, due to the high
sensitivity to air and low stability under the electron beam
of the alanates, electron microscopy studies have been lim-
ited. Some work has been done using scanning electron
microscopy (SEM) to study the morphology of Ti doped
Li3AlH6 [7]. The microstructure of NaAlH4 containing TiF3
additive direct after the ball milling process and after 15 cy-
cles has been studied recently using transmission electron
microscopy (TEM) and SEM [15].
In this work, powder X-ray diffraction (PXD), TEM and
SEM in conjunction with energy dispersive X-ray analysis
(EDS) have been used to study undoped LiAlD4. We have
examined the different stages in the decomposition process
with respect to microstructure and phases present. Studies
of material mixed with Ti/V have been reported elsewhere
[9,16].
decomposition was stopped at 185 ◦C for R1 and 275 ◦C for
R2.
PXD data were collected in a flat plate geometry with an
INEL MPD diffractometer equipped with a curved CPS 120
position sensitive detector covering a 2θ-range of from 1 to
◦
˚
120 , λ = 1.540562 A (Cu K␣1 radiation). The powder was
uniformly spread in a thin layer on the sample holder and
then covered by a thin plastic film to prevent the samples
from reacting with oxygen or moisture during the measure-
ments. Rietveld refinements were carried out using the Ri-
etica program [17]. Structural data for LiAlD4 and Li3AlD6
were taken from [10] and [18], respectively. Voigt profile
functions were used and the background was modelled by
Cheby II polynomials.
Electron microscopy samples were prepared in the glove
box by mechanical grinding of the powder in a silicate cru-
cible. The dry powder was spread on a holey carbon film
supported on a copper grid. Samples were transferred into
the TEM using a vacuum container and exchanged into the
TEM by means of a removable glovebag device, which main-
tained an Ar overpressure during the transfer process. For
SEM, a simple sealed transfer container has been developed
that allows the sample to be placed directly into the SEM
air-lock entry chamber, while still under an inert Ar atmo-
sphere. Electron microscopy studies were performed with a
Jeol 2010F TEM operated at 200 kV and a Hitachi S-4300SE
SEM operated at 10 kV.
2. Experimental
LiAlD4 (Sigma–Aldrich; >95% chemical purity, >98%
isotope purity) was obtained as powder and stored under Ar
atmosphere. Commercial LiAlH4 powders contain a signifi-
cant amount of impurities and in order to avoid this problem,
LiAlD4 was selected for the experiments. All handling of the
samples was done in Ar atmosphere in a glove box to pre-
vent reaction with moisture and oxygen in the air. The sample
contained traces of LiCl (about 0.2 wt.%).
Thermal desorption spectroscopy (TDS) experiments
were carried out in dynamic vacuum at a heating rate of
2 ◦C/min. The background vacuum level, with empty sample
holder, was approximately 2 × 10−6 mbar. In order to study
the decomposition reaction at a given stage, the sample holder
was removed from the furnace at the specific temperature
during heating and quenched into cold water. The decompo-
sition curves after R1 and after R2 are shown in Fig. 1. The
3. Results and discussion
3.1. X-ray diffraction
Fig. 2 shows the diffraction patterns of the LiAlD4 dur-
ing the decomposition process, determined from Rietveld
refinements. The phase compositions determined from quan-
titative phase analysis are indicated in Table 1. Before
Fig. 1. Decomposition curves from TDS experiments after R1 and R2. The heating rate was 2 ◦C/min.

C.M. Andrei et al. / Journal of Alloys and Compounds 395 (2005) 307–312
309
Fig. 2. PXD patterns of LiAlD4 before decomposition and at 185 and 275 ◦C (R1 and R2), respectively. The positions for the Bragg peaks are given for LiAlD4,
LiCl, Li3AlD6 and Al from top to bottom. A peak from the plastic film covering the sample is indicated by (ꢀ).
Table 1
amount of remaining deuterium was not detectable in the
TDS experiment.
Phase composition in mol% of the compound before its decomposition and
after R1 and R2
Before
decomposition
After R1
After R2
3.2. TEM studies
LiAlD4 (mol%)
Li3AlD6 (mol%)
Al (mol%)
99
–
–
–
27
72
1
–
16
82
2
A TEM bright field image of the LiAlD4 before the de-
composition process is shown in Fig. 3a. The corresponding
selected area diffraction (SAD) pattern is shown in Fig. 3b.
The nano-sized grains are randomly oriented inside the pow-
der particle, as indicated by uniform rings in the diffraction
pattern. The d-spacings are listed in Table 2. Comparison
LiCl (mol%)
1
The LiD phase was not possible to detect by X-ray.
decomposition, LiCl impurity was the only additional crys-
talline detectable phase in the material. For the sample pre-
pared at 185 ◦C (after R1) no traces of LiAlD4 were detected.
The relative phase composition of Li3AlD6 and Al indicates
that some Li3AlD6 also has been decomposed at this tempera-
ture. For X-rays, LiD is nearly impossible to distinguish from
Al, since Al has a significant stronger X-ray scattering than
Li and both compounds crystallize in the same space group,
Table 2
Index of the d-spacings of the LiAlD4 sample before the decomposition
process
˚
˚
d-Spacing (A),
TEM
PXD data (A)
Identification
h k l
2.64
2.40
1.84
1.76
1.65
1.41
1.34
2.667
2.414
1.832
1.760
1.647
1.410
1.348
LiAlD4
LiAlD4
LiAlD4
LiAlD4
LiAlD4
LiAlD4
LiAlD4
0 2 2
1 0 2
1 3 3
1 2 4
2 0 2
1 5 1
1 3 5
Fm3m with almost the same lattice parameter, aAl = 4.04 and
˚
a
LiD = 4.06 A. The TDS data (Fig. 1) indicate that R2 should
be completed at 275 ◦C. However, the X-ray data (Fig. 2 and
Table 1) show that the R2 sample still contains a significant
amount of Li3AlD6 (about 50%). The release of the small
Fig. 3. (a) TEM bright field image, (b) SAD and (c) dark field image of a LiAlD4 particle before the decomposition process.

310
C.M. Andrei et al. / Journal of Alloys and Compounds 395 (2005) 307–312
with the PXD data shows that all the rings correspond to
the LiAlD4 phase. The TEM dark field image of the LiAlD4
(Fig. 3c) is made from reflections indicated in the SAD pat-
tern. The measured grain size of the LiAlD4 phase for the
TEM images varies between 5 and 70 nm. The calculated
grain size from the width of the diffraction peaks by the
Scherrer equations for LiAlD4 phase was about 90 nm. The
discrepancy in the measured grain size between TEM and
PXD can be due to the detection limit of the X-ray diffrac-
tometer.
The sample was comparatively stable under the elec-
tron beam and high-resolution (HR) experiments could be
performed. A HRTEM image of the crystalline part of
a particle before the decomposition process is presented
˚
in Fig. 4. The measured lattice spacing of 2.43 A corre-
sponds to d102 = 2.414 A of the LiAlD4 phase from the PXD
data.
˚
Fig. 4. HRTEM image showing lattice fringes in the LiAlD4 sample before
the decomposition process. The indicated lattice plane are indexed as LiAlD4
(1 0 2).
A bright field image of the sample, after reaction R1, is
presented in Fig. 5a. Indexing of the SAD pattern (Fig. 5b)
shows that metallic Al spots have the strongest intensity and
˚
Table 3
only one reflection with the measured d-spacing of 1.64 A
Index of the d-spacings of the sample after reaction 1 (comparison of SAD
and PXD)
was clearly identified as the Li3AlD6 phase. Other interplanar
distances from Li3AlD6 are too similar to those from Al and
too weak to be distinguished by electron diffraction. All the
d-spacings are confirmed from the PXD data. The main d-
spacings measured in TEM compared with the PXD data are
listed in Table 3.
The same experiments were performed on the alanate par-
ticles after R2. A bright field image and the corresponding
SAD are presented in Fig. 6a and b. As described in the
X-ray section, LiD and Al show d-spacings that are indis-
˚
TEM d (A)
Li3AlD6 (PXD)
Al (PXD)
˚
˚
d (A)
h k l
d (A)
h k l
2.32
2.05
1.64
1.45
1.23
0.93
2.330
2.051
1.640
1.456
1.238
0.930
2 0 2
1 2 3
2 4 0
0 2 5
2 3 5
6 1 1
2.338
2.024
1 1 1
2 0 0
1.431
1.221
0.928
2 2 0
3 1 1
3 3 1
Fig. 5. (a) Bright field image and (b) SAD of a particle after reaction 1.
Fig. 6. (a) Bright field image and (b) SAD of a particle after reaction 2.

C.M. Andrei et al. / Journal of Alloys and Compounds 395 (2005) 307–312
311
Table 4
Index of the d-spacings of the alanate heated to 275 ◦C (after R2) (comparison of SAD and PXD)
˚
TEM d (A)
Li3AlD6 (PXD)
Al (PXD)
LiD (PXD)
˚
˚
˚
d (A)
h k l
d (A)
h k l
d (A)
h k l
2.84
2.31
2.02
1.65
1.47
1.23
2.816
2.330
2.051
1.652
1.456
1.238
0 2 2
2 0 2
1 2 3
3 2 1
0 2 5
2 3 5
2.338
2.024
1 1 1
2 0 0
2.348
2.033
1 1 1
2 0 0
1.431
1.221
2 2 0
3 1 1
1.438
1.226
2 2 0
3 1 1
Fig. 7. (a) Secondary electron (SE) image and (b) backscattered electron (BSE) image of LiAlD4 before the decomposition process.
tinguishable by diffraction. All inter-planar distances d of
3.3. SEM studies
these two phases overlap (see Table 4). Reflections corre-
˚
sponding to the 0 2 2 planes (d = 2.816 A) and 3 2 1 planes
Fig. 7a shows a SEM image taken with secondary elec-
trons (SE), giving information about the three-dimensional
morphology of the particle. The corresponding image taken
with backscattered electrons (BSEs) (Fig. 7b), which is more
sensitive to changes in the composition, shows uniform con-
trast, which is consistent with the presence of a single phase
LiAlD4.
˚
(d = 1.652 A) of the Li3AlD6 phase can be clearly identified
in the diffraction pattern. Other spacings for this phase are
too close to the Al spacings to allow them to be distinguished.
The identification of Li3AlD6 in the electron diffraction data
for the sample heated to 275 ◦C is in agreement with the PXD
results.
Fig. 8. (a) Secondary electron (SE) image and (b) backscattered electron (BSE) image of the sample after reaction R1.
Fig. 9. (a) Secondary electron (SE) image and (b) backscattered electron (BSE) image of the alanate after reaction 2.

312
C.M. Andrei et al. / Journal of Alloys and Compounds 395 (2005) 307–312
By comparison, after the first reaction, the morphology of
the R1 reaction into Al particles distributed over the surface
of the Li containing phases. After the second reaction (R2)
there seem to be a more mixed stage. The presence of Al at
the surface of the alanates after the R1 reaction confirmed
the long range transporting mechanism observed earlier by
in-situ X-ray diffraction [19].
theparticlesischangedasseenintheSEimageof Fig. 8a. Dif-
ferences in the fine scale composition can be seen in the BSE
image of Fig. 8b. The finer, bright particles corresponding to
higher average atomic number are found to give a strong Al
EDS signal, while the rest of the cluster corresponds to the
Li3AlD6 phase. Close examination of the SEM images sug-
gests that Al particles are distributed over the surface of the
alanate. In-situ X-ray diffraction studies of the decomposed
NaAlH4 [19]reportedtheformationofaluminiumcrystallites
with sizes larger than 100 nm at the surface of the alanates. A
long-range mechanism transporting Al from a decomposition
site within the alanate particles to the sites where the growth
of Al crystallites occurs was reported. This is consistent with
our observations.
Acknowledgement
The Norwegian Research Council is gratefully acknowl-
edged for the financial support on the project “Hydro-
gen storage in metal hydrides” contract number 135502/
431.
The SEM SE image of the alanate after R2 taken is shown
in Fig. 9a. By comparison, in the corresponding BSE, fine
scale variations in the composition are shown in Fig. 9b.
The fine distribution of the Al particles on the surface of
the alanate in the sample after R1 is not clearly observed in
these images, although some brighter contrast is seen in BSE
image. The reason for this is not fully understood. Contrast in
the image is also due to LiD phase, which is not detectable by
EDS. Further in-situ SEM experiments using a heating stage
can clarify this.
References
[1] G.D. Berry, S.M. Aceves, Energy Fuels 12 (1) (1998) 49.
[2] B. Bogdanovic, M. Schwickardi, J. Alloys Compd. 253 (1997) 1.
[3] A.E. Finholt, A.C. Bond Jr., H.I. Schlesinger, J. Am. Chem. Soc. 69
(1947) 1199.
[4] W.E. Garner, E.W. Haycock, Proc. R. Soc. (Lond.) A 211 (1952)
335.
[5] J. Block, A.P. Gray, Inorg. Chem. 4 (1965) 304.
[6] N.N. Mal’tseva, A.I. Golovanova, Russ. J. Appl. Chem. 73 (2000)
747.
[7] J. Chen, N. Kuriyama, Q. Xu, H.T. Takeshita, T. Sakai, J. Phys.
Chem. B 105 (2001) 11214.
[8] V.P. Balema, J.W. Wiench, K.W. Dennis, M. Pruski, V.K. Pecharsky,
J. Alloys Compd. 329 (2001) 108.
4. Conclusions
[9] D. Blanchard, H.W. Brinks, B.C. Hauback, P. Norby, Mater. Sci.
Eng. B 108 (2004) 54.
[10] B.C. Hauback, H.W. Brinks, H. Fjellva˚g, J. Alloys Compd. 346
(2002) 184.
[11] H.W. Brinks, B.C. Hauback, P. Norby, H. Fjellva˚g, J. Alloys Compd.
351 (2003) 222.
[12] V. Mikheeva, S. Arkhipov, Zh. Neorg. Khim. 12 (1967) 2025.
[13] T.N. Dymova, V.N. Konoplev, D.P. Aleksandrov, A.S. Sizareva, T.A.
Silina, Koord. Khim. 21 (1995) 175.
[14] J. Dilts, E. Ashby, Inorg. Chem. 11 (1972) 1230.
[15] C.M. Andrei, J.C. Walmsley, H.W. Brinks, R. Holmestad, C.M.
Jensen, B.C. Hauback, Appl. Phys. A, in press, doi:10.1007/s00 339-
004-3106-Z.
[16] C.M. Andrei, J. Walmsley, H.W. Brinks, R. Holmestad, D. Blan-
chard, B.C. Hauback, G.A. Botton, J. Phys. Chem. B, in press.
[17] B. Hunter, Commission For Powder Diffraction, International Union
of Crystallography, Newsletter No. 20 (1998).
[18] H.W. Brinks, B.C. Hauback, J. Alloys Compd. 354 (2003)
143.
The microstructure of LiAlD4 has been studied at differ-
ent stages of the decomposition process using a combination
of PXD, TEM, SEM and EDS. Before decomposition, PXD,
HRTEM and SAD confirmed the LiAlD4 phase with low level
of LiCl impurity. After the first desorption reaction, metal-
lic Al gave the stronger intensity in the diffraction pattern.
Inhomogeneous distribution of the phases was confirmed in
SEM BSE images with Al being present largely at the sur-
face of the alanate particles. After the second reaction, the
interplanar distances of the Al and LiD phases overlap, and
this made their identification difficult. Two SAD reflections
correspond unambiguously to the Li3AlH6 phase. The PXD
data confirmed also the existence of the Li3AlH6 phase in
the material. The presence of this phase after R2 reaction
was unexpected. The SEM images during the decomposi-
tion process showed the phase separation. The homogeneous
structure observed before decomposition is transformed after
[19] K.J. Gross, S. Guthrie, S. Takara, G. Thomas, J. Alloys Compd. 297
(2000) 270.